Virtual world simulation represents one of the hottest fields in information technology. In virtual world simulation, clients receive 3D scene data associated with the virtual world from a server, and then rendering of the received data is performed locally, at the client. However, the computation required for rendering is usually overwhelming such that, for example, for a high fidelity movie image, only a few images can be rendered in a day. To remedy this problem, certain technology employs pre-rendering methods, i.e., pre-rendering shot images before actually playing them. However, unlike movie images that are pre-shot, virtual world images are generated in real time and change unpredictably, which in turn, requires the client to possess high computational power. For example, the client is required to have an auxiliary device such as a dedicated graphical processor to perform real-time rendering, which of course, greatly increases the costs of the client. Additionally, lightweight computational devices, such as laptops, which are currently very popular, command low computational power making real-time rendering in such devices extremely difficult or impossible.